Overproduction and Characterization of the Bacillus subtilis Anti-sigma Factor FlgM*

FlgM is an anti-sigma factor of the flagellar-specific sigma (ς) subunit of RNA polymerase inBacillus subtilis, and it is responsible of the coupling of late flagellar gene expression to the completion of the hook-basal body structure. We have overproduced the protein in soluble form and characterized it. FlgM forms dimers as shown by gel exclusion chromatography and native polyacrylamide gel electrophoresis and interacts in vitro with the cognate ςDfactor. The FlgM·ςD complex is a stable heterodimer as demonstrated by gel exclusion chromatography, chemical cross-linking, native polyacrylamide gel electrophoresis, and isoelectric focusing. ςD belongs to the group of sigma factors able to bind to the promoter sequence even in the absence of core RNA polymerase. The FlgM·ςD complex gave a shift in a DNA mobility shift assay with a probe containing a ςD-dependent promoter sequence. Limited proteolysis studies indicate the presence of two structural motifs, corresponding to the N- and C-terminal regions, respectively.

FlgM is an anti-sigma factor of the flagellar-specific sigma () subunit of RNA polymerase in Bacillus subtilis, and it is responsible of the coupling of late flagellar gene expression to the completion of the hook-basal body structure. We have overproduced the protein in soluble form and characterized it. FlgM forms dimers as shown by gel exclusion chromatography and native polyacrylamide gel electrophoresis and interacts in vitro with the cognate D factor. The FlgM⅐ D complex is a stable heterodimer as demonstrated by gel exclusion chromatography, chemical cross-linking, native polyacrylamide gel electrophoresis, and isoelectric focusing. D belongs to the group of sigma factors able to bind to the promoter sequence even in the absence of core RNA polymerase. The FlgM⅐ D complex gave a shift in a DNA mobility shift assay with a probe containing a D -dependent promoter sequence. Limited proteolysis studies indicate the presence of two structural motifs, corresponding to the N-and C-terminal regions, respectively.
In bacteria genes are regulated mainly at the transcriptional level. In addition to the cis-acting promoter sequences, a large number of activators and repressors modulate initiation of transcription. Furthermore initiation depends on alternative sigma factors that direct the RNA polymerase to specific promoter sequences. The activity of sigma factors may itself be regulated, either by stability as is the case for the Escherichia coli stationary phase S (1) or by interaction with anti-sigma factors (2). Anti-sigma factors have been characterized for B and F of B. subtilis (3)(4)(5) and for the flagellar dedicated sigma of Salmonella (6,7) and 70 of E. coli upon infection with the bacteriophage T4 (8). Anti-sigma factors have also been described in Pseudomonas aeruginosa and Myxococcus xanthus (9 -11). The activity of the anti-sigma factors is exercised by specific protein-protein interaction, even though the details of the molecular mechanisms involved are still largely to be elucidated (12)(13)(14)(15).
FlgM was first identified in Salmonella as a negative regulator of the flagellar-specific sigma factor 28 , also called FliA (6,7), and later shown to interact specifically with the cognate sigma factor, forming a stable complex (15). FlgM has also been described in B. subtilis, where it appears to accomplish a similar regulatory function, coupling the early and late flagellar gene expression (16 -18). The action of FlgM is thought to occur by specific interaction with the late flagellar sigma factor D .
The B. subtilis FlgM has now been overproduced in Escherichia coli, purified, and shown to form a complex with purified D protein. Limited proteolysis studies suggest that FlgM contains a relatively stable N-terminal domain.
To purify FlgM, BL21 (DE3) (pBG9) cultures were grown at 37°C with shaking in 150 ml of Luria broth medium supplemented with 1% (w/v) glucose and 100 g/ml ampicillin. At A 600 of 3-4, the cells were collected by centrifugation and resuspended in 1 liter of Luria broth medium with 100 g/ml ampicillin; isopropyl-␤-D-thiogalactopyranoside was added to 0.1 mM, and growth was continued for 3 h at 21°C, until an A 600 of 0.7. The cells were recovered by centrifugation and resuspended in 100 ml of A buffer (20 mM NaH 2 PO 2 , 29.6 mM Na 2 HPO 2 , 50 mM NaCl, 0.1 mM EDTA, pH 8.0, 10% (v/v) glycerol). Cells were lysed by sonication, and the lysate was centrifuged to remove cell debris and inclusion bodies. The supernatant, containing the soluble fraction of the protein, was loaded on a FPLC 1 (Pharmacia) S-Sepharose 26/10 column, pre-equilibrated with A buffer. The FlgM protein was purified by elution with a 0 -0.5 M gradient of NaCl in A buffer; 19 mg of FlgM were recovered, eluting with 0.3 M NaCl. The protein was further purified by gel exclusion chromatography on an FPLC Sephacryl S-100 column in B buffer (10 mM Tris-HCl, pH 7.5, 100 mM NaCl, 10 mM MgCl 2 , 1 mM EDTA, pH 8.0, 10% (v/v) glycerol).
Overexpression and Purification of B -B was overexpressed by using a pT7-5 derivative, kindly supplied by W. G. Haldenwang. For purification we followed the procedure described by Huang et al. (20) for W. The last steps (heparin-Sepharose and Superdex) were omitted and substituted by a single HiTrap Q column (Amersham Pharmacia Biotech).
Physical Characterization of FlgM-The molecular mass of FlgM and its fragments were determined by matrix-assisted laser desorption and ionization time-of-flight mass spectroscopy using a matrix of sinapinic acid or of ␣-cyano-4-hydroxycinnamic acid. The N-terminal amino acid sequences were determined by automated Edman degradation using an Applied Biosystem 477A liquid pulse protein sequencer.
FlgM⅐ D Binding Assay-Purified proteins FlgM and D were incubated together at equimolar ratio at 0°C for 1 h, and the mixture was * This work was supported in part by grants from Ministero dell'Università e della Ricerca Scientifica e Tecnologica and the Consielio Nazionale delle Ricerche Target Project on Biotechnology. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
§ Supported by a fellowship from International Centre for Genetic Engineering and Biotechnology (Trieste, Italy). Permanent address: Genetic Engineering and Biotechnology Center, Havana, Cuba.
Chemical Cross-linking-Purified proteins were cross-linked in a 50 l reaction mixture containing 50 mM HEPES, pH 7.5, 100 mM NaCl, 10 mM MgCl 2 , 10% glycerol by adding ethylene glycol-bis(succinic acid N-hydroxysuccinimide ester) (EGS, Sigma) or dithio-bis(succinimidyl propionate) (DSP, Sigma) to a final concentration of 1 mg/ml. EGS and DSP were predissolved in dimethyl sulfoxide at 10 mg/ml. After incubation at 0°C for 2 h, L-lysine was added to a final concentration of 20 mM. Cross-linked products were separated and analyzed by SDS-PAGE.
Isolelectric Focusing-Isoelectric focusing was performed in nondenaturating conditions, using horizontal electrophoresis apparatus with 1 M NaOH and 1 M H 3 PO 4 as electrode solutions. The samples (5-20 l) were applied to a 1-mm-thick gel using isoelectric focusing applicator (Pharmacia). After focusing, proteins were visualized by Coomassie Blue staining.
To analyze the protein components associated with the shifted DNA band, the corresponding portion of the gel was cut out, rerun in 15% SDS-PAGE, and transferred to a nitrocellulose filter. The filter was probed with anti-FlgM polyclonal antibodies, followed by a secondary antibody conjugated with horseradish peroxidase. The peroxidase activity was visualized with an enhanced chemiluminescence kit (ECL, Amersham Corp.) Limited Proteolytic Digestion-FlgM (2.5 mg/ml) was incubated at 21°C for 30 min in reaction containing 20 mM HEPES, pH 7.5, 50 mM KCl, 10 mM MgSO 4 , and 0.01 mg/ml subtilisin. Reactions were stopped by adding 10 mM phenylmethylsulfonyl fluoride and visualized by 20% SDS-PAGE. Samples were blotted to polyvinylidene difluoride membrane and analyzed by automated amino acid sequencing. In a second experiment the products of limited proteolysis were separated by reverse phase HPLC (Jasco on an Aquapore RP-300 C 8 column (25 ϫ 0.46 cm). The solvent system used was either 0.1% (v/v) trifluoroacetic acid in H 2 O (system A) or 0.075% trifluoroacetic acid in CH 3 CN (system B). The elution was with a linear gradient from 0 to 50% system B in system A, in 70 min at a flow rate of 0.7 ml min Ϫ1 . The elution of peptides was monitored at 220 nm. Aliquots were analyzed by matrixassisted laser desorption and ionization time-of-flight mass spectroscopy and Edman degradation.

Overproduction and Purification of FlgM-The B. subtilis
FlgM is a 9.8-kDa protein (16). The sequence coding for FlgM was obtained by polymerase chain reaction amplification of B. subtilis chromosomal DNA, using primers with a NdeI and BamHI site, respectively; after purification and restriction, flgM was cloned in the expression vector pET12a. The coding sequence was followed by the original stop codon, and the expected product was only FlgM, without fusion to any resident sequence. The resultant plasmid, pBG9, was transformed into E. coli BL21 (DE3). Following induction with isopropyl-␤-Dthiogalactopyranoside expression of FlgM was obtained at high level, but in the insoluble protein fraction. After several attempts we found conditions at which the majority of FlgM was in the soluble fraction. The volume of the culture was 1 liter in a 5-liter volume flask, the incubation temperature was 21°C, and the concentration of the isopropyl-␤-D-thiogalactopyranoside inducer 0.1 mM. The soluble fraction was purified by S-Sepharose chromatography, followed by a gel exclusion column of Sephacryl S-100. After this step, the protein was estimated to be at least 95% pure by SDS-PAGE and Coomassie staining (Fig. 1). The final yield was 19 mg of FlgM/liter of culture of E. coli grown to an A 600 of 0.7.
Characterization of FlgM-N-terminal determination by Edman degradation of an aliquot of the purified protein gave the sequence MKINQFG corresponding to the first seven amino acids of FlgM as deduced from the nucleotide sequence (16). The mass, measured by mass spectrometry was 9,931 Da, in agreement with the estimated mass of 9,862 Da. We used gel exclusion chromatography to determine whether FlgM is a dimer in solution. The elution volume on a Sephacryl S-100 column corresponds to a molecular mass of about 21 kDa, suggesting that FlgM is a dimer. This was further confirmed by PAGE under nondenaturing conditions (Fig. 2). The dimer nature of FlgM is also corroborated by results of chemical cross-linking experiments. Incubation of FlgM with crosslinker generated a faster migrating species, probably due to internal cross-linking, and a band of approximately 20 kDa, again suggestive of a homodimer (Fig. 3A, lane 3). At least one other anti-sigma factor, SpoIIAB of B. subtilis, has been reported to be present as a dimer (4,21). We have no knowledge of the possible oligomerization of FlgM of Salmonella.
FlgM Forms a Complex with D -FlgM is an anti-sigma factor specific for D as shown by in vitro and in vivo experiments (17). 2 To investigate the interaction of the two proteins, we purified D according to Chen and Helmann (23). Upon mixing and incubation at 0°C for 1 h, the two proteins formed a complex that could be resolved by gel exclusion chromatography on a Superose 12HR column. The complex was stable even in 0.6 M NaCl, suggesting a participation of hydrophobic 2 D. Barillà and M. Chamberlin, personal communication. interaction between the two polypeptide chains in the complex.
The purified complex was treated with the homobifunctional cross-linking agents DSP and EGS. Analysis of the DSPtreated product by SDS-PAGE showed a single band of approximately 40 kDa (Fig. 3A). Similar results were obtained upon treatment with EGS (data not shown). No bands corresponding to the two proteins, FlgM and D , were observed, suggesting that the complex is very stable and long lived, because it is not in equilibrium with the monomers. The same cross-linked product was obtained by treating the two proteins, immediately upon mixing. FlgM and D treated separately with DSP did not produce a band of the same mobility as that obtained when the mixture was treated (Fig. 3A, lanes 3 and 5).
The complex appears to be an heterodimer by the mass of the cross-linked complex estimated by SDS-PAGE and electrophoresis of purified complex under nondenaturing gel gradient electrophoresis. In addition the mass of the EGS cross-linked complex was measured by mass spectrometry, and the observed value of 43,399 Da is in good agreement with the theoretical value of 39,311 Da (9, 862 Da of FlgM ϩ 29, 449 Da of D ), considering the presence of an unknown number of EGS molecules. The same type of heterodimeric structure has been reported for the Salmonella FlgM. 28 complex (15). The B. subtilis purified complex, not treated with the cross-linking agent, had a pI of about 6.8 as determined experimentally by isoeletric focusing; distinct from the observed pI of D (5.9) and the deduced pI of 9.9 of FlgM (data not shown).
The complex gave a single band in PAGE under nondenaturing conditions (Fig. 3B, lane 5). Using antibodies anti-D to detect free D in Western blotting, in nondenaturing conditions, we estimated that the apparent dissociation constant for the FlgM⅐ D complex is 10 Ϫ6 M or less (data not shown).
To evaluate the specificity of the complex formation, we performed similar experiments with FlgM and a different sigma factor of B. subtilis, B , which is known to interact with the cognate anti-sigma, RsbW (3). We failed to observe any interaction between the two proteins (Fig. 3B, lanes 2 and 3). (In the electrophoretic system used, FlgM is not visible, because it moves out of the gel.) The FlgM⅐ D Complex Binds DNA-D belongs to the group of sigma factors devoid of region 1, responsible for masking the rest of the molecule. Thus these factors can bind to the promoter sequence even in absence of core RNA polymerase (23).
FlgM inhibits initiation of transcription from D -dependent promoters in vitro. 2 Unexpectedly the FlgM⅐ D complex gave a supershift in a gel retardation assay with a probe containing the flagellin, D -dependent promoter PD-6 (22). Not only was the probe retarded more than observed with only D , but the amount of probe bound to the complex was much higher. The experiment was performed in the presence of cold unspecific plasmid DNA. To show that the supershift was indeed due to the complex, a gel slice corresponding to the shifted band was subjected to SDS-PAGE and analyzed by Western immunoblotting with anti-FlgM antibodies. These polyclonal antibodies cross-react with D and allowed the detection of the presence of both proteins in the shifted band (Fig. 4). The same shift was observed when the probe was first incubated with D and an equimolar amount of FlgM added 5 min prior to the electrophoresis (Fig. 4, lane 5). In our tests, D alone gave a weak shift of the probe (Fig. 4, lane 4), and this is probably due to incomplete renaturation of the protein solubilized by guanidine hydrochloride. The higher amount of probe retained in the presence of FlgM suggests that the anti-sigma factor may facilitate the renaturation of D .
Partial Proteolysis of FlgM-To gain information on the structural organization of FlgM, we performed experiments of limited proteolysis. Partial digestion with subtilisin gave consistently one major product that in PAGE had a mobility of approximately 5 kDa (Fig. 5A). The same electrophoretic pattern was obtained following incubation at 21°C for at least one h (data not shown). Edman degradation of the material transferred to polyvinylidene difluoride membranes, indicated the presence of the N terminus of FlgM. In a parallel experiment the products of partial proteolysis were subjected to separation by HPLC. Only one peak was observed, but two N-terminal sequences were obtained in approximately equimolar amounts:  MKINQFGTQSVNPY and IENGSYKVDANHIA (Fig. 5B). The first sequence corresponds to the N terminus of FlgM, whereas the second one corresponds to residues 65-78 of the protein.
Various attempts to further purify the two components were unsuccessful. The same fraction obtained by HPLC was analyzed by mass spectrometry and gave the values of 5754 and 2284 Da, respectively (Fig. 5B). The M r of 5754 is in good agreement with the calculated value (5753 Da) of a peptide consisting of the first 51 residues of FlgM; the smaller peptide of M r 2284, should correspond to amino acids Ile 65 to Phe 84 (calculated mass, 2279 Da) toward the C terminus of the protein. A minor component of M r 5810 was also observed as a shoulder of the main peak of M r 5754; we infer that the proteolytic cleavage can occur either upstream or downstream of Gly 52 .
The results of the partial proteolysis are summarized in the Fig. 5C. We conclude that FlgM is organized in two regions: the N-terminal part (residues 1-51) and the C-terminal portion (residues 65-84), which appear structured and resistant to proteolytic cleavage. The two structural motifs are presumably connected by a loop, easily accessible to cleavage. DISCUSSION We have overexpressed and obtained in soluble form the anti-sigma factor FlgM of B. subtilis, the main regulator of D activity. The protein is a dimer in solution, as determined by gel exclusion chromatography and PAGE electrophoresis under nondenaturing conditions. The anti-sigma factor FlgM forms a stable and long lived complex with the cognate sigma factor D . Highly stable complexes have been reported for two other B. subtilis sigma-anti-sigma couples, SpoIIAB⅐ F and RsbW⅐ B (24,25).
FlgM⅐ D complex still binds to the promoter sequence, but the fact that FlgM does not interfere with the interaction of D at the promoter site may have a marginal physiological significance, as already noted for the DNA binding by the same sigma factor (23). Therefore, if, as it seems plausible, in vivo the promoter recognition is performed by the holoenzyme, the specific anti-transcriptional activity of FlgM could be exerted in three ways: (i) sequestering of free D molecules in the FlgM⅐ D complex, (ii) removing D from the holoenzyme, and (iii) hampering a productive interaction between the holoenzyme and the promoter.
In the case of bacteriophage T 4 , the anti-sigma AsiA forms a tight complex with the 70 subunit (8,26). Nevertheless a detailed analysis of its mechanism of action suggests that AsiA interacts with the holoenzyme, thereby modifying the interaction between the enzyme and promoter sequence and not simply dissociating the sigma factor from the core (27)(28)(29). Re-cently it has been shown that the Salmonella FlgM, in addition to sequestering the free sigma factor, binds to the holoenzyme and increases the rate of dissociation of sigma from the core (30).
The anti-sigma factors themselves are regulated. In the case of SpoIIAB and RsbW, the sigma factor is set free to accomplish transcription by the action of an anti-anti-sigma factor (SpoI-IAA and RsbV, respectively), which under certain conditions forms a complex with the anti-sigma protein (25,(31)(32)(33).
As for FlgM we do not yet know the mechanism regulating its activity. A clue derives from work on the Salmonella homologue of FlgM that has been shown to be secreted into the medium by the flagellar secretion system (34,35). When the flagellar hook-basal body complex is assembled, the devoted class III secretory apparatus export FlgM from the cell, thereby relieving the inhibition exerted by the anti-sigma factor. In solution the free Salmonella FlgM is mostly unfolded, which may facilitate its secretion; it forms a stable complex with 28 , but upon binding, the C-terminal half of FlgM becomes structured, with the N terminus still unfolded and thus potentially competent for secretion (15). The present results from partial proteolysis indicate a different organization of the B. subtilis FlgM polypeptide. The N-terminal region is structured, at least as deduced from partial proteolysis and not unfolded as suggested for the Salmonella FlgM from NMR studies. The structural differences may hint at different mechanisms in the way to regulate the anti-sigma factor activity in the cell.